Research On Automatic Surface Inspection Technology For Scraping Workpiece

Scraping technology is one of basic key technologies of mechanical engineering. Nowadays, scraping processing and scraping quality inspection are still mainly done manually around the world. The inspection process is cumbersome with a large amount of work, and the inspection results could be easily influenced by many human factors which severely restrict the scraping quality and efficiency. Therefore, the scraping process automation is a kind of future trend. The automatic surface inspection technology is one of key problem in auto-scraping research, the research production of the technology will greatly promote the development of auto-scraping thereby expedite advanced manufacturing process.Based on the newest domestic and foreign achievements of machine vision technologies, combining knowledge of fitter, modern manufacturing technology and control technology all together, the automatic surface inspection system is established. And the following achievements are obtained:A geometric correction method based on a plane image is presented. The method can correct the perspective and lens distortions real-timely in an image without knowing the posture and position of the camera. The method’s principle is simple and its function is easy to realize. Experiments are performed on different images and the results of other methods(Zhengyou Zhang’s calibration method and camera calibration toolbox for Matlab) are compared with these results, the comparison results showing that this method has a higher efficiency and precision.A high-points recognition and localization method of scraped surface based on background subtraction is proposed. The method can separate foreground objects from the rest of the image, recognize and locate the high-points. In the method, different methods of image gray processing are discussed, and the optimal method is determined. All kinds of different noise type are analyzed, linear filtering and morphological filtering are combined together to filter and reduce noise. The existing scraping technologies are analyzed, the effect of circularity threshold and rectangle degree threshold on the high-points recognition success rate is studied, based on the above research, optimum threshold of high-points recognition is determined, high-points recognition and localization algorithm is proposed. Chinese scraping quality testing standard is studied, a reasonable evaluation index of scraping surface is established to ensure the accuracy and feasibility of inspection.Based on OpenCV and wxWidgets, automatic surface inspection software for scraping workpiece using C++ language is developed. Quality inspection of the scraped workpiece is done, the experimental result is given, and the result will play a guiding role to our in-depth scraping process.